Feature/few shot learning

[HeerakKashyap]

🚀 Add Comprehensive Few-Shot Learning Support (Issue #237)

📋 Overview

This PR implements comprehensive few-shot learning capabilities for igel, addressing GitHub issue #237 "Add Support for Few-Shot Learning". The implementation includes meta-learning algorithms, domain adaptation utilities, and transfer learning capabilities.

✨ Features Added

🧠 Meta-Learning Algorithms

• Model-Agnostic Meta-Learning (MAML): Complete implementation with inner/outer loop optimization
• Prototypical Networks: Embedding-based few-shot learning approach
• Configurable parameters: Learning rates, task sampling, adaptation steps

�� Domain Adaptation

• Fine-tuning adaptation: Adapt entire models to target domains
• MAML-based adaptation: Use meta-learning for domain transfer
• Support for source-to-target domain transfer

�� Transfer Learning

• Feature extraction: Extract features from pre-trained models
• Fine-tuning capabilities: Adapt pre-trained models to new tasks
• Easy integration with existing models

🛠️ CLI Commands

• igel few-shot-learn - Train few-shot learning models
• igel domain-adapt - Perform domain adaptation
• igel transfer-learn - Apply transfer learning

📁 Files Added/Modified

New Files

• igel/igel/few_shot_learning.py - Core few-shot learning module
• examples/few_shot_learning_example.yaml - Configuration example
• examples/few_shot_learning_demo.py - Comprehensive demonstration script
• tests/test_few_shot_learning.py - Complete test suite
• docs/few_shot_learning.md - Detailed documentation

Modified Files

• igel/igel/data.py - Added few-shot learning models to models_dict
• `igel/